You'll still generate just as much heat. Remember, the laws of thermodynamics are:
1. You can't win. 2. You can't even break even.
If you convert X watts of solar power into electricity, then that same X watts will eventually be converted into heat. Some in transmission losses, some in actual use, but it'll all end up as heat. Really.
Yes, greenhouse gasses are a good blanket. Like I said, they insulate the Earth, and that's something to look at. But you can't tell me that producing many gigawatts of heat doesn't add to the problem.
If you harness solar, it doesn't hit the ground, but you still use the energy, and it still ends up as heat. That heat may be produced somewhere where the electrical energy is converted to some other form, but it's still going to end up as heat. You can't escape it.
I don't believe that nuclear power is entirely green. It may not produce greenhouse gasses or other pollutants, but there's one more thing:
If we're worried about global warming, the amount of heat that we produce should be a factor as well as how much we insulate the Earth with greenhouse gasses. Nuclear power releases very large amounts of heat that would otherwise not be released.
On the other hand, wind, solar, and hydro power merely harness energy that is going to end up as heat anyway.
So, if you're worried about heating up the Earth, would you rather release gigawatts of extra heat, or would you rather harness gigawatts of heat that's being released anyway?
You can pick the worst three, but it doesn't apply to the rest. If I used your logic, I'd point to Switzerland, and say that the obvious choice is to issue a machine gun to every adult.
The long and short of it is that if people want to fight and kill, they're going to fight and kill. I've been in "weaponless" countries, where there was as much (or more) killing than in the countries you mention. You can't stop violence with laws, you can only disarm the law-abiding people who are NOT the ones you should be worrying about.
Move to a country that will let you carry a pistol, too.
steve
Where many people miss the point...
on
Is Swap Necessary?
·
· Score: 4, Insightful
People like to claim that swap can always improve performance, by swapping out unused sections of memory, allowing for more memory to throw at apps or disk cache.
Well, *most* apps won't just arbitrarily consume memory, so endless amounts of memory won't help. And disk cache gets you greatly diminishing returns.
One of the machines I use has 3 gigs of memory. It will swap out unused programs, in an attempt to free up more memory. The joke is that it simply can't use all three gigs. After half a year of uptime, there's still over half a gig completely unused, because the apps don't take memory, and there's not that much to put in disk cache.
Obviously, that's a pathological case. And there are pathological cases at the other extreme. But as memory prices keep dropping over the long run, swap does become less and less useful.
You're missing the fact that a lot of radio stations use transmitters that pump out 50,000 watts or more. With that kind of power, there's a reason why your radio reception is so darn good - not to mention the fact that it's a one-way transmission.
Still, drive through a thick-walled underground parking garage, lose your radio signal. Do you whine and moan? Not at all. But if your puny cell phone with a 1-watt transmitter can't penetrate ten meters of concrete and dirt, you whine like it was the end of the world.
I haven't seen any maps of cellular coverage, but I've been with Verizon for years. A lot of my friends have used AT&T, Cricket, Voicestream, T-Mobile, and a couple of others. So far, I've never seen one that got as good of coverage as I do. (I should mention that I have a good-quality tri-mode phone.)
Last summer, we travelled through Illinois and D.C.. My wife and I never had a problem with our phones, coverage, or signal strength. Everyone else, not using Verizon, had nothing but dead time and problems through the entire trip. One of them had even signed up with AT&T specifically to get good coverage while travelling.
Alright, now I see where I went wrong. It wasn't a quad-board with two sets of memory sockets, it's a dual-board that only has memory sockets hooked to one of the processers, the K8W.
I could swear that some company was showing off a quad-Opteron board that will fit in an ATX form factor - achieved by leaving the memory slots off of two of the processers. Halving your memory bandwidth to save some real estate is a tough compromise for me to accept, but I guess there are applications where you wouldn't incur much of a performance penalty.
Not me. I happen to have a dual Athlon, and it's got plenty of CPU cycles, but it's limitted to a single 64-bit, 266-MHz memory subsystem. The dual Athlon would have two 128-bit 333- (or 400-) MHz subsystems.
Now, if someone would make a dual-Athlon chipset with a 128-bit memory controller, that might change things - but we both know that's not going to happen.
The cheapest I could find an FX53 (2.4 GHz) was $733. An Opteron 150, the same thing (except for the locked multiplier) was over $100 cheaper.
When I bought an AthlonXP, I got the mobile version for the unlocked multiplier. It was an extra $15. But an extra $120 for the privelige? You've got to be awfully excited to pay that much.
Even so, it's still insane. The fact that it will take multiple 3GHz+ cores to let you do the same tasks you can do right now with a single 1GHz core is just silly.
It's like saying "Oh, it's not a problem that this car has an extra two tons of useless weight. After all, we can just throw in one of those huge V10's!"
Doom3, however, is fully multi-threaded in order to support SMP systems. That means that the games which license the Doom3 engine will be multithreaded. And it also means that anyone who wants to challenge Id will also have to step up to the plate.
Earlier today, I heard someone moaning that the need to support multiple processers was useless baggage that would pull down the video game industry. Quite the contrary, increases in computing performance have always helped the video game industry, and the ability to tap into two processers instead of one is another way that they can increase their use of yoru CPU cycles.
There are a *lot* of simultaneoush things happening in a video game - in addition to the rendering and sound, you've also got to handle AI for a good number of characters, and physics for a (usually) large number of objects, and those are two things that can chew up CPU cycles.
Sure, it takes some work and intelligence to get all of the code to work together. But that's alright, that's how things have been improving for a long time now.
Right now I have a dual AthlonMP 1800+ system at work. Even though the case itself gets plenty of ventilation, the air coming out of the power supply is niiiiiice and toasty warm. Even though I'm quite happily married, it's still nice that the attractive girls from the office come over on chilly days to sit in my chair and put their feet on the top of my computer!
Actually, back in 1992, the computer I really wanted was $10,000. It was a 486DX/33 with 64 megs of RAM, 1 gigabyte hard disk, and a 1 gigabyte tape backup - unbelievably huge at the time.
I'm soooooooo glad I never bought one.
One of my coworkers in about 1994 had spent some thousands of dollars on a motherboard with 64 megabytes of sram as the main memory. Insanely fast at the time. But again, I'm soooooooo glad I didn't buy one.
I've got a120mm NMB fan that pushes air at over 60 CFM, and you have to have your ear within a foot of it to even be able to hear it. One of those on the back of a small form factor case, blowing in, through, and out strategically placed slots would be far more than enough to keep it within an allowable temperature range.
Just so long as they actually give you a set of memory slots for each chip. Some companies (Tyan) have put out quad-boards that only have memory slots for two of the chips. It'll work, and it saves a lot of real estate, but then you're completely losing one of the greatest strengths of the Opterons.
Whenever my friends from warmer climates come and visit, they always ask why we don't have raised studs on the roads here. When I point out to them that we have to plow snow off of our roads on a regular basis through the winter, they see that the raised studs would only last until the first snowfall. : )
I'll do it for $2,000, and get to pocket more than half of that!
"Thermoelectrics". You know, peltiers. I just picked up a couple of high-quality, surplus 45-watt Peltiers for $4 each. Each being able to pump up to 30 watts of heat, it certainly wouldn't take many of them to keep that beer keg nice and chilly.
So, just for grins, let's say that I buy 20 of those, and run them at much less than full power. That keeps each spot from getting so hot that it needs a heat sink. That's what, $80 so far? Then, it's a matter of building the power supply/temp controller and sewing up a wrap.
Out of this guy's $20,000 grant, he'll probably be able to spend about $19,500 on "restocking vital supplies" (refilling all of the kegs he's drunk).
4096 Itanium2's, pumping out 20 teraflops, and it's supposed to be the fastest? I thought that Red Storm (10,000+ Opterons) was pumping out around 40 teraflops/second.
My mail server's been chugging along for about 4 years now, and is terrifically reliable. So, I turned off the fsync() calls, so things like that don't really matter any more, as the kernel's disk cache can do what it was designed to do. Throughput went up by more than a factor of ten.
Some day, a fan, power supply, or UPS will die. But getting 10x the performance for 4+ years justifies losing the two minutes worth of email that wasn't flushed to disk when that day comes.
I use a Sony 40x burner with buffer underrun, and choose the "optimal" setting. If I'm burning a large ISO, by the end of the disk, it'll be sustaining 34x to 38x, and out of many hundreds of disks, I've got a *100%* success rate.
Slashdot Mirror
You'll still generate just as much heat. Remember, the laws of thermodynamics are:
1. You can't win.
2. You can't even break even.
If you convert X watts of solar power into electricity, then that same X watts will eventually be converted into heat. Some in transmission losses, some in actual use, but it'll all end up as heat. Really.
steve
Yes, greenhouse gasses are a good blanket. Like I said, they insulate the Earth, and that's something to look at. But you can't tell me that producing many gigawatts of heat doesn't add to the problem.
steve
If you harness solar, it doesn't hit the ground, but you still use the energy, and it still ends up as heat. That heat may be produced somewhere where the electrical energy is converted to some other form, but it's still going to end up as heat. You can't escape it.
steve
I don't believe that nuclear power is entirely green. It may not produce greenhouse gasses or other pollutants, but there's one more thing:
If we're worried about global warming, the amount of heat that we produce should be a factor as well as how much we insulate the Earth with greenhouse gasses. Nuclear power releases very large amounts of heat that would otherwise not be released.
On the other hand, wind, solar, and hydro power merely harness energy that is going to end up as heat anyway.
So, if you're worried about heating up the Earth, would you rather release gigawatts of extra heat, or would you rather harness gigawatts of heat that's being released anyway?
steve
You can pick the worst three, but it doesn't apply to the rest. If I used your logic, I'd point to Switzerland, and say that the obvious choice is to issue a machine gun to every adult.
The long and short of it is that if people want to fight and kill, they're going to fight and kill. I've been in "weaponless" countries, where there was as much (or more) killing than in the countries you mention. You can't stop violence with laws, you can only disarm the law-abiding people who are NOT the ones you should be worrying about.
steve
Move to a country that will let you carry a pistol, too.
steve
People like to claim that swap can always improve performance, by swapping out unused sections of memory, allowing for more memory to throw at apps or disk cache.
Well, *most* apps won't just arbitrarily consume memory, so endless amounts of memory won't help. And disk cache gets you greatly diminishing returns.
One of the machines I use has 3 gigs of memory. It will swap out unused programs, in an attempt to free up more memory. The joke is that it simply can't use all three gigs. After half a year of uptime, there's still over half a gig completely unused, because the apps don't take memory, and there's not that much to put in disk cache.
Obviously, that's a pathological case. And there are pathological cases at the other extreme. But as memory prices keep dropping over the long run, swap does become less and less useful.
steve
You're missing the fact that a lot of radio stations use transmitters that pump out 50,000 watts or more. With that kind of power, there's a reason why your radio reception is so darn good - not to mention the fact that it's a one-way transmission.
Still, drive through a thick-walled underground parking garage, lose your radio signal. Do you whine and moan? Not at all. But if your puny cell phone with a 1-watt transmitter can't penetrate ten meters of concrete and dirt, you whine like it was the end of the world.
steve
I haven't seen any maps of cellular coverage, but I've been with Verizon for years. A lot of my friends have used AT&T, Cricket, Voicestream, T-Mobile, and a couple of others. So far, I've never seen one that got as good of coverage as I do. (I should mention that I have a good-quality tri-mode phone.)
Last summer, we travelled through Illinois and D.C.. My wife and I never had a problem with our phones, coverage, or signal strength. Everyone else, not using Verizon, had nothing but dead time and problems through the entire trip. One of them had even signed up with AT&T specifically to get good coverage while travelling.
steve
What the ...?
Alright, now I see where I went wrong. It wasn't a quad-board with two sets of memory sockets, it's a dual-board that only has memory sockets hooked to one of the processers, the K8W.
I could swear that some company was showing off a quad-Opteron board that will fit in an ATX form factor - achieved by leaving the memory slots off of two of the processers. Halving your memory bandwidth to save some real estate is a tough compromise for me to accept, but I guess there are applications where you wouldn't incur much of a performance penalty.
steve
Not me. I happen to have a dual Athlon, and it's got plenty of CPU cycles, but it's limitted to a single 64-bit, 266-MHz memory subsystem. The dual Athlon would have two 128-bit 333- (or 400-) MHz subsystems.
Now, if someone would make a dual-Athlon chipset with a 128-bit memory controller, that might change things - but we both know that's not going to happen.
steve
The cheapest I could find an FX53 (2.4 GHz) was $733. An Opteron 150, the same thing (except for the locked multiplier) was over $100 cheaper.
When I bought an AthlonXP, I got the mobile version for the unlocked multiplier. It was an extra $15. But an extra $120 for the privelige? You've got to be awfully excited to pay that much.
steve
Even so, it's still insane. The fact that it will take multiple 3GHz+ cores to let you do the same tasks you can do right now with a single 1GHz core is just silly.
It's like saying "Oh, it's not a problem that this car has an extra two tons of useless weight. After all, we can just throw in one of those huge V10's!"
steve
You're pretty much right, but soon won't be.
Doom3, however, is fully multi-threaded in order to support SMP systems. That means that the games which license the Doom3 engine will be multithreaded. And it also means that anyone who wants to challenge Id will also have to step up to the plate.
Earlier today, I heard someone moaning that the need to support multiple processers was useless baggage that would pull down the video game industry. Quite the contrary, increases in computing performance have always helped the video game industry, and the ability to tap into two processers instead of one is another way that they can increase their use of yoru CPU cycles.
There are a *lot* of simultaneoush things happening in a video game - in addition to the rendering and sound, you've also got to handle AI for a good number of characters, and physics for a (usually) large number of objects, and those are two things that can chew up CPU cycles.
Sure, it takes some work and intelligence to get all of the code to work together. But that's alright, that's how things have been improving for a long time now.
steve
Right now I have a dual AthlonMP 1800+ system at work. Even though the case itself gets plenty of ventilation, the air coming out of the power supply is niiiiiice and toasty warm. Even though I'm quite happily married, it's still nice that the attractive girls from the office come over on chilly days to sit in my chair and put their feet on the top of my computer!
steve
Actually, back in 1992, the computer I really wanted was $10,000. It was a 486DX/33 with 64 megs of RAM, 1 gigabyte hard disk, and a 1 gigabyte tape backup - unbelievably huge at the time.
I'm soooooooo glad I never bought one.
One of my coworkers in about 1994 had spent some thousands of dollars on a motherboard with 64 megabytes of sram as the main memory. Insanely fast at the time. But again, I'm soooooooo glad I didn't buy one.
steve
I've got a120mm NMB fan that pushes air at over 60 CFM, and you have to have your ear within a foot of it to even be able to hear it. One of those on the back of a small form factor case, blowing in, through, and out strategically placed slots would be far more than enough to keep it within an allowable temperature range.
steve
To imitate Kahn from King of the Hill...
"Haha! Joke's on you!"
The FX is nothing more than a 1xx Opteron (for the dual-channel memory controller) that's relabeled as an overpriced Athlon64.
steve
Just so long as they actually give you a set of memory slots for each chip. Some companies (Tyan) have put out quad-boards that only have memory slots for two of the chips. It'll work, and it saves a lot of real estate, but then you're completely losing one of the greatest strengths of the Opterons.
steve-O
The torrent page has links for the -test3 line, but I don't see any for FC2-final. Am I missing something?
steve
Whenever my friends from warmer climates come and visit, they always ask why we don't have raised studs on the roads here. When I point out to them that we have to plow snow off of our roads on a regular basis through the winter, they see that the raised studs would only last until the first snowfall. : )
Steve
I'll do it for $2,000, and get to pocket more than half of that!
"Thermoelectrics". You know, peltiers. I just picked up a couple of high-quality, surplus 45-watt Peltiers for $4 each. Each being able to pump up to 30 watts of heat, it certainly wouldn't take many of them to keep that beer keg nice and chilly.
So, just for grins, let's say that I buy 20 of those, and run them at much less than full power. That keeps each spot from getting so hot that it needs a heat sink. That's what, $80 so far? Then, it's a matter of building the power supply/temp controller and sewing up a wrap.
Out of this guy's $20,000 grant, he'll probably be able to spend about $19,500 on "restocking vital supplies" (refilling all of the kegs he's drunk).
steve
4096 Itanium2's, pumping out 20 teraflops, and it's supposed to be the fastest? I thought that Red Storm (10,000+ Opterons) was pumping out around 40 teraflops/second.
Did I miss something?
steve
My mail server's been chugging along for about 4 years now, and is terrifically reliable. So, I turned off the fsync() calls, so things like that don't really matter any more, as the kernel's disk cache can do what it was designed to do. Throughput went up by more than a factor of ten.
Some day, a fan, power supply, or UPS will die. But getting 10x the performance for 4+ years justifies losing the two minutes worth of email that wasn't flushed to disk when that day comes.
steve
12X to get a 99% chance? Wow.
I use a Sony 40x burner with buffer underrun, and choose the "optimal" setting. If I'm burning a large ISO, by the end of the disk, it'll be sustaining 34x to 38x, and out of many hundreds of disks, I've got a *100%* success rate.
steve